UltraWide-Band (UWB) low-power devices are becoming an increasingly effective solution for tracking and positioning systems, especially in IoT devices with limited power sources. Background models can be created by analysing the radio frequency or channel impulse response in a static environment. Comparing this background model with runtime measurements it is possible to identify moving objects inducing changes in the impulse responses, and as a consequence, we can compute the distance of the object from the radar devices, usually consisting of one transmitter and one receiver. In particular, using multiple receivers, it becomes possible to elaborate a 2D position of the moving object. This type of tag-less passive tracking system has the advantage of retrieving the trajectory of a moving object without the adoption of any additional electronic device. In this paper, we present an IoT positioning architecture achieving an RMSE of about 30 cm.